|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
| lizmat | and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/12/24/...ars-later/ | 00:11 | |
| AlexDaniel | please, #moarvm, take a look at github.com/rakudo/rakudo/issues/2564 | 00:36 | |
| AlexDaniel looks up | 00:37 | ||
| oh :) | |||
|
02:02
lizmat left
03:53
reportable6 left,
shareable6 left,
squashable6 left,
evalable6 left
03:54
shareable6 joined
03:55
p6bannerbot sets mode: +v shareable6,
squashable6 joined,
evalable6 joined
03:56
p6bannerbot sets mode: +v squashable6,
p6bannerbot sets mode: +v evalable6
03:57
reportable6 joined
03:58
p6bannerbot sets mode: +v reportable6
05:41
shareable6 left
05:43
shareable6 joined
05:44
p6bannerbot sets mode: +v shareable6
07:36
shareable6 left
07:38
shareable6 joined
07:39
p6bannerbot sets mode: +v shareable6
|
|||
| nine | timotimo: that's exactly what I have done ;) But I'm no longer sure its result is valid... | 09:21 | |
|
09:21
robertle left
09:22
robertle joined
09:23
p6bannerbot sets mode: +v robertle
09:26
domidumont joined
09:27
p6bannerbot sets mode: +v domidumont
|
|||
| timotimo | say, are those int attributes 64bit sized inside the Match object? | 09:45 | |
| but if !cursor_fail is broken, it would have exploded way earlier | 09:46 | ||
| so yeah, no clue what's up with that | |||
| nine | I guess the only way to find out is to disable spesh optimizations until the error goes away. But that will take longer than I have before going to my parents' for Christmas | 09:57 | |
| I've already tried disabling the getattr/bindattr optimizations in P6Opaque. That's why I suspect the bisect to be misleading | 09:59 | ||
| timotimo | a patch to make that would probably be a single line to main.c :) | 10:03 | |
| nine | Err....I meant disabling optimizations one by one until we find the broken one | ||
| timotimo | oh | 10:04 | |
| yeah that'd take a bit longer | |||
| AFKBBL | |||
|
10:04
lizmat joined,
p6bannerbot sets mode: +v lizmat
12:59
domidumont left
13:12
brrt joined,
p6bannerbot sets mode: +v brrt
|
|||
| brrt | \o | 13:12 | |
| nine++ # debugging | |||
| timotimo | i'm sure the bytecode compiler could be loads faster if it didn't calculate 3*4 every insert_bytecode call :P | 13:51 | |
| hm, how do we get an int from inside a BOOTInt in nqp code? | 13:56 | ||
| nqp-m: my $foo := nqp::list; nqp::push($foo, 1); my int $foo := nqp::shift($foo); say($foo) | 13:57 | ||
| camelia | Redeclaration of symbol $foo at line 2, near "; say($foo" at gen/moar/stage2/NQPHLL.nqp:819 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:panic) from gen/moar/stage2/NQP.nqp:2903 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/nqp.moarvm:vari… |
||
| timotimo | nqp-m: my $foo := nqp::list; nqp::push($foo, 1); my int $bar := nqp::shift($foo); say($bar) | ||
| camelia | 1 | ||
| timotimo | nqp-m: my $foo := nqp::list; nqp::push($foo, 1); my int $bar := nqp::unbox_i(nqp::shift($foo)); say($bar) | ||
| camelia | 1 | ||
| timotimo | elems, coerce_in, const_n64, div_n, coerce_in, coerce_iu, trunc_u16 | 14:09 | |
| amazing | |||
|
14:12
evalable6 left
14:14
evalable6 joined
14:15
p6bannerbot sets mode: +v evalable6
14:16
brrt left
|
|||
| timotimo | smrt_numify may want to be one of the things that counts as an unbox | 14:17 | |
| could put an nqp::unbox_i around basically every method call in write_frame. that would be pretty terrible :\ | 14:27 | ||
| because at the moment it's numifying the integer objects that these methods return, and then coerces that resulting num to an int | 14:33 | ||
| but tbh i haven't measured the performance yet | 14:34 | ||
| so it could very well be wasted time :) :) | |||
|
14:39
eater left
14:40
domidumont joined
14:41
p6bannerbot sets mode: +v domidumont
15:08
domidumont left
15:41
eater joined,
p6bannerbot sets mode: +v eater
|
|||